摘要
将Tian差值扩展技术应用于彩色图像,提出一种基于预测误差差值扩展(PEDE)和余数调整的彩色图像可逆数据隐藏算法。针对传统差值扩展技术存在过分修改像素灰度值、须嵌入定位图等缺点,首先利用色彩分量间的相关性减小差值,并将差值扩展量分散到2个色彩分量中;其次,由2个色彩分量中像素的预测平均值决定可用于扩展嵌入的差值,对不能用于扩展嵌入的差值则用可逆对比图RCM变换嵌入数据,无需保存溢出定位图;最后,对直方图平移技术进行改进,实现嵌入容量和失真控制,提取端在提取信息时可无损地恢复原始图像。实验结果表明,与其他算法相比,本文算法在同等嵌入率下可取得更好的图像质量,算法复杂度更低。
A reversible data hiding algorithm is proposed for color images using prediction error difference expansion and RCM transform. To avoid the drawback of Tian＇s algorithm which must embed a location map and the quality decline of stego-images from excessive modulation to pixels, this algorithm uses cor- relation of color components to decrease difference and disperse smaller expansion to two components, the expand embedding difference is determined with average of prediction values,and RCM transform is used to embed data in other difference, so the algorithm does not need location map. Finally, the histo- gram shifting technique is improved to control capacity and distortion. Experimental results show that the quality of stego-image is significantly improved, and the complexity is lower compared with other new or classical algorithms.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2010年第4期572-578,共7页
Journal of Optoelectronics·Laser
基金
国家民委重点科研项目基金资助(MZY02004)
关键词
可逆数据隐藏
预测误差差值扩展
(RCM)变换
失真控制
溢出定位图
reversible data hiding
prediction error difference expansion
RCM transform
distortion con-trol
overflow location map
